Slashdot Mirror


Market Share Reports On Linux

spizkapa writes: "IDC has predicted that Linux will grow steadily along with Microsoft in the near future in the home PC (client) market, as well as including numbers that prove Linux's acceptance rate is fantastic. " The numbers look nice, especially in the server area, but it's too bad that things weren't broken down more. I'm also wondering where *BSD fits in -- I assume under UNIX, but it's unfortunate that they weren't broken out separately.

26 of 204 comments (clear)

  1. Re:They're missing something though... by embobo · · Score: 3

    I have the answer! Convince all distro makers to send a message to install-counter@example.com everytime the distro is installed. Suppose the installed machine has no 'net connection? Simple: require one! Disable the sw unless the email is sent (and an encypted and signed receipt using the ethernet card mac addr or some other guid of the machine to prevent spoofing) within 30 days of install. What's that? Its being done already? Well then, time to innovate!

    Have a cron job send an email to current-counter@example.com with the current time in the subject once a week. Then we'll know how many running linux boxes there are at any given week.

    I'll be damned if I'm going to run linux until I _know_ that there are at least 4.5 million boxes running it. My enterprise MIS CIO CTO MBA training has made me very smart, truly. Sure, I read Information Week. I see the one page ads with linux in 72pt type. But hell, I'm no fool. I need an official report from a $5000/yr newsletter, complete with facts and figures, before I switch my multi-trillion dollar dot-com from Windows 2000 ME Data Center SP4.1 to linux.

    This conversation can serve no pupose anymore. Thank you, please drive through.

  2. Re:They're missing something though... by Matts · · Score: 3

    See the latest netcraft addendum to the stats though... They found that most of those boxes are Linux+Apache boxes running massive virtual hosting. In fact register.com is responsible for (I can't recall the exact figure, but something line...) 20% of those numers, and they are just hosting default templates saying this site hasn't been uploaded yet. While in a way that's good for Linux+Apache ("Linux can easily host thousands of virtual domains"), it doesn't give any indication of the number of unique machines running linux. And neither does the number of unique IP addresses for similar reasons.

    --

    Matt. Want XML + Apache + Stylesheets? Get AxKit.
  3. Re:They're missing something though... by mattdm · · Score: 3
    Not so. If a business-type guy wants to sell software, he is (or should be, at least) concerned with the number of people who are actually running Linux. OS sales figures might be interesting for people thinking of marketing a Linux distro, but not for someone who wants to sell software that runs on Linux.

    --

  4. Maybe there should be a mechanism in distros. by Hanno · · Score: 3

    Sometimes I think that there should be an anonymous counting mechanism included with the distributions.

    Similar to pine's anonymous user counter (for those who don't know about it: The first a new user starts pine on a fresh system, it asks you wether it may send an anonymous mail to the pine developers, just so that they have a rough idea about their user base).

    Or similar to Debian's popularity contest (it reports the list of installed packages to the developers, so that they can see what is used most-often and thus deserves additional work over a rarely used software package).

    Just imagine if every Linux distribution would do something like this after its first run of the installation:

    "Have your new machine counted! May I send a one-time message to the Linux user counter for you?"

    It might even be an incentive for the distributions to do so if the user share of the different flavours of Linux were counted that way.

    ------------------

    --

    ------------------
    You may like my a cappella music
  5. don't worry about this kind of data by jetson123 · · Score: 3
    Obviously, these numbers are strongly biased in favor of Windows: most installations of Linux never show up in them, and many "shipments" of Windows (OEM licenses, site licenses, etc.) never actually get used. IDC's numbers probably also underestimate the Mac user base.

    But why worry about it? Companies that make business decisions based on flawed data are likely to fare poorly in the marketplace in the long run. Just think of the relative costs and reliability of a Windows-based web site and a UNIX based equivalent.

  6. Re:They're missing something though... by swb · · Score: 3
    Also, shipments of an operating system that can be installed a theoretically infinite number of times are obviously skewed when compared to OSs like NT, which are to be installed on a single server.
    I'll agree with you that box sales is a lousy way to sample linux adoption compared with other OS which are SUPPOSED to be installed only once per copy.

    NT's previous competitor in the LAN server market was Novell Netware, and Netware was super strict about licensing -- if you installed it on your network, you installed it once and it had rigid user limits. NT had the honor system, which lets people install it more than once. I think that NT's growth in the LAN server market can be partially attributed to the "honor system" licensing approach. And I'm curious how many shops take this to conclusions they probably aren't supposed to.
    "It's just temporary until we get the new hardware."

    "The IT staff were evaluating a new product and didn't want to do it on the production system."

    "We just needed a new box really quick. I think the legit software has been ordered."

  7. Re:They're missing something though... by Mullen · · Score: 3

    Exactly!

    Even suff like the Linux User Registration (or whatever it is called) is not useful for all known installs of Linux.
    It is too easy download an ISO image of your favorite Linux dist and install where ever you want.

    I had an extensive discussion with one of my old Statatics professors in college. After explaining what Linux was, how you can get it for free, how there is no central Linux download site, how there are so many places to get Linux, how there are no restrictions on installing Linux, and how easy it is share the same CDROM, NFS, or FTP site; she came to the conclusion that there is no way to count the machines that have Linux on it.
    There are simply too many variables in the "Linux counting problem."

    So my point is, when you read something about how many Linux machines there are out there, the author is smoking crack. It is nearly impossiable to count all the Linux machines in the world.

    --
    Linux O Muerte!
  8. Re:Linux's relative growth by SoftwareJanitor · · Score: 3

    Exchange isn't even a reason to run NT anymore. HP's OpenMail is a replacement for Exchange that is open source, more cost effective, and runs on a better platforms, with a choice of platforms (Linux and many commercial UNIXes). And yes, Windows 9x and NT desktop users can use Outhouse... Errr... LookOut... Err Outlook as their mail client if they want. Although to be fair, it is really Outlook that is more responsible for the security problems with that system than Exchange.

  9. Re:Linux's relative growth by SoftwareJanitor · · Score: 3

    But you don't know me. I'm not interested in changing that, either.

    I'm not that interested in changing that either. What you don't say in order to interest me is any reason why I should care. I never said Linux was for everyone. Being for anyone is different than being for everyone. There is no product that is for everyone, and no product should try to be. Products that try to be everything to everyone inevitably end up being self-limited to being mediocre at best.

    At any rate, as far as I can tell, the number of backsliders like yourself is far outweighed by the number of people going the other direction.

    I've noticed for quite some time now that Linux advocates always try to dis W2K any time they can.

    Most of that I think comes from being fed up with Microsoft. I spend less than 10% of my time (less and less lately) dealing with Microsoft products, and that causes 90% of the frustration I have. There are very few OSes that I've used that I haven't grown to like more with use. The only two things I can think of that have gone the other way have been VMS and MS-DOS/Windows/NT.

    At any rate Microsoft and their apologists (paid and otherwise) also try to 'dis' anything that isn't Microsoft any time they can. That is just the way it goes. Ask yourself this -- if Microsoft was so great, why do they get so much negative reaction these days? If Linux had no merits, why would so many people be lining up behind it even though they often have no financial interest in doing so?

    It's gonna eat their lunch in the end.

    We will have to agree to disagree on this one. While I won't venture to say that Linux will ever rule the entire world the way that Microsoft has, that isn't a bad thing. In fact, what I really want isn't necessarily a world without Microsoft, it is a world in which NO single company or technology rules everything. Something, someday, may manage to unseat Linux from its place on my machines, but it sure won't be Windows 2000, and its highly unlikely that it will be any of its successors. Maybe one of the *BSDs, maybe something totally new.

    In the end, Microsoft will implode due to its own gravitational force or break up from within or gradually succumb to outside pressure from various competitors. It is inevitable that every empire will fall.

    Hopefully Linux will be one of the things that brings Microsoft down a notch or three, but hopefully it won't be the only thing.

  10. Re:Linux's relative growth by AJWM · · Score: 3

    I know I shouldn't feed the troll, but...

    why your toy OS is better than Windows.

    That's the real joke. I and many others have been calling Windows, and it's underlying DOS, a "toy OS" for years. Think of all the things Windows can't do that a real OS can -- like protect itself from a renegade app, or permit multiple, different simultaneous users, or just manage to stay up for a few months without a reboot.

    Now, I'll grant that NT, being based on VMS, was more of a real OS -- but even VMS was never the Unix killer that DEC had hoped. And that's was ... more and more application level stuff (the GUI? come on!) has been crammed into the NT 'kernel', to its detriment.

    No serious observer of operating systems considers Linux to be anything less than a real OS, nor Windows (9x) any more than a toy. A few will grudgingly grant NT 'real' status, with caveats.

    On second thoughts...

    The above troll question can simply be answered:

    Because it will run an IBM S/390. Some "toy".

    --
    -- Alastair
  11. Home market growing with Windows sounds right... by Blackwulf · · Score: 3

    Just about everybody that I have spoken to who has Linux installed, has it dual-booted with Windows. Because of this, the number will show them growing the same, because many casual users (not all) have both the Windows and a Linux OS on the same machine.

    These numbers don't show what's used more, it's just showing what's been bought and installed.

  12. Licenses by alehmann · · Score: 3
    Linux has seen a dramatic increase in new license shipments growing from 15.8% of server operating environment shipments in 1998 to 24.4% in 1999, absorbing much of the expansion of the market.

    Since the GPL is freely transferable, am I the only one who doesn't think that license shipments have anything to do with Linux growth or sales?

  13. Re:They're missing something though... by AugstWest · · Score: 3

    Well, our CEO is definitely a business-type guy. We're building a portal, and hsoting some applications in-house. You should hear him on the phone talking about Linux as the greatest thing since sliced bread....

    Why is it important to the biz-types? Well, it's because of the price tag.

    The free downloads and installs are very important to biz guys, just not in this context.

    Which of course, you weren't denying, and this isn't an argument with your points, just kind of a further clarification.

  14. Re:Distortion by alexhmit01 · · Score: 3

    No, businesses make decisions with money. Assessing the financial strength of different players also uses money. This is measuring percentage of server license sales.

    Who cares how many servers are deployed? It is irrelevant to EVERYONE. Your decision to use an OS should be completely independent of what other people are doing, right?

    Numbers matter to zealots and brats who want to scream and yell.

    Money, however, matters. It makes a lot of BIG differences. If I am a interested in the revenue model of a OS, then the money does matter.

    Business analysis is not aimed for slashdot, it is aimed towards business executives. Ya know, the people who worry about this for a living, not a pissing contest.

    Alex

  15. Re:They're missing something though... by rgmoore · · Score: 3
    They're couting shipments of Linux, which totally misses out on all of the downloads.

    Yeah, this is an important point. They give Linux 4.1% of 98.8 million client shipments and 24% of 5.7 million server shipments. That suggests that Linux is now shipping more units for use as desktops than for servers.

    The question, though, is whether that accurately represents the usage patterns. I'd expect that a lot of server farms either use downloads or a single install disk for multiple machines. At the same time, I'd expect a lot of home user/hobbyist types (like me) to try out multiple different versions of Linux to find the best distribution. Heck, you can get CD's for $2 or less from places like Linux Mall, and I have probably 10 different versions that I've bought either there or as the full distribution, while I have only 3 computers with Linux installed. When you look at the revenue from Linux (apparently less than $100 million on about 5 million copies sold) that suggests that a lot of the copies are the very low cost ones.

    --

    There's no point in questioning authority if you aren't going to listen to the answers.

  16. Bullshit polling methods by AndyChrist · · Score: 3

    Biased polling is a time-honored tradition! It's the american way! Do the polling exclusively at slashdot. Announce to the world that Linux has 50 percent (or whatever) market share!

  17. Distortion by joel.neely · · Score: 3

    Notice how often these types of analyses measure "market share" in terms of dollar volume. This creates a built-in bias against Linux, FreeBSD, etc. "If you're inexpensive -- let alone free -- you must not count for much..."

  18. A prettier graph by tagishsimon · · Score: 3

    one of the graphs from the IBM page mantioned earlier today seems apropos (and prettier...)

  19. DUH by Kalrand · · Score: 3

    It only stands to reason that as the market for PC's continues to explode, the number of operating system licences will expand as well.

    I could have told you that the number of people using any major OS would increase so long as new PC's are being sold.

    duh!

    Kalrand

    -the voice of reason

  20. These numbers are bogus! by Ergo2000 · · Score: 3

    The title is a humor title, however it's interesting seeing the reaction to anything like this. Okay folks, Linux users make up 90% of the population! There, are you happy? Is that more "real" to you?

    Shipments do mean nothing. There are countless hoardes of people out there who never paid for their copy of Windows, and by the same token there are tonnes of people who've picked up countless Redhat x.x CDs at the local bargain bin but have never done more than put them on their rack.

    All that really matters at the metrics of actual usage, because copies sitting on people's shelves mean nothing. By that token the most recent study put Linux users at approximately 0.29% of the Internet browsing public, down from 0.32% of a month earlier. I'm sure that will be a rather stark number for a lot of the rose coloured dreamers lusting for the day that Linus is the true leader of the masses.

  21. Not a clue about Apple by Otter · · Score: 4
    Note that this paragraph is discussing clients, not servers.
    The upcoming Mac OS X (read "10") represents a complete overhaul and a radical strategy for Apple (for one thing, core OS X source code is freely available). Before birthing it, Apple aborted its Copland and Rhapsody OS plans and cut loose its MkLinux team after years of laboring on these projects, losing time and credibility. However, MkLinux (now independent) and LinuxPPC are shipping and can run on older Macs.

    Huh?
    • OS X is Rhapsody, albeit a lot later and a lot different than originally promised.
    • I hardly think resources spent MkLinux were what was holding Apple back. That was, what, three full-time engineers? How much time was invested in MkLinux, let alone credibility? And the work on Mach was valuable for OS X.
    • And, of course, OS 9 is still shipping and is at least as well suited for office desktop use than MkLinux or LinuxPPC.
  22. IDC is not reliable. by BitGeek · · Score: 4

    I used to work in the industry and had numerous discussions with "analysts" from IDC, Forrester, Jupiter, etc. My experience there showed me that all of these "Reports" and "Analysis" are really propaganda under another form.

    The analysts are paid to write reports showing the growth and projected growth of various industry segments. These reports are commisioned by companies in the industries in question. They are then sold to other companies in the smae industry.

    For instance, say I'm a major software vendor, and want to do a new product in the project management software industry- I'll have IDC write a report justifying why this industry is going to grow at %20 a year for the next decade. They will go and look for supporting evidence, but since statistics can often lie, when you go looking for a specific growth rate, you will find it.

    So, for instance, while Apple continues to hold about %30 of the installed base of computers, IDC shows them with only %5 of the market, because its convenient for them to only count new computer sales for Apple. While Linux is probably 5 times as popular as they show in their "survey" windows is shown to be dominant.

    Why? Because the people who have the money who pay IDC have a vested interest in windows.

    This is no different than the investment bank that brought acompany public issuing a report with a buy recommendation on that companies stock.

    There is NO credibility here, whatsoever.

    --
    Yeah, and you guys panned the ipod too: http://apple.slashdot.org/article.pl?sid=01/10/23/ 1816257
  23. Meanwhile, back in the real world... by Black+Parrot · · Score: 4
    See Netcraft's analysis of OS market shares in their July 2000 survey.

    In short, and using the conservative numbers:
    • Linux : 29.99%
    • Windows (all types) : 28.32%
    • Solaris : 16.33%
    • Other : 23.59%
    • Unknown : 1.76%
    These numbers are for Web servers only; they filter out duplicate names for the same site, and also "placeholder" sites, but do not filter out virtual hosting. So it's "sites" rather than "machines".

    They also used some statistical sampling, but do not report a margin of error.

    --
    --
    Sheesh, evil *and* a jerk. -- Jade
  24. Re:They're missing something though... by SoftwareJanitor · · Score: 5

    I'll be damned if I'm going to run linux until I _know_ that there are at least 4.5 million boxes running it.

    Netcraft says that there are over 10 million servers running Apache on the publicly accessable internet. Those are machines that are running basically 24/7, not 'dual boot' machines or people's dialup boxes that are getting miscounted. Netcraft also says that over 1/3 of the Apache servers are running Linux as their OS. According to the Linux counter, less than 30% of Linux boxes are used as web servers.

    Given those numbers, how could anyone reasonably believe that there aren't well more than 4.5 million Linux boxes running out there?

    And yes, I am probably just feeding the trolls...

  25. They're missing something though... by AugstWest · · Score: 5

    They're couting shipments of Linux, which totally misses out on all of the downloads.

    I've set up a good 30 linux boxes in my time, and I've never purchased a single copy.

    Once again, Linux doesn't fit into the normal boxes used to judge these things. The distribution model for it is entirely unheard of, and so they don't have any mechanisms for couting this massively popular means of obtaining linux.

    Also, shipments of an operating system that can be installed a theoretically infinite number of times are obviously skewed when compared to OSs like NT, which are to be installed on a single server.

  26. Comparing Apples and Oranges by martyb · · Score: 5

    Interesting article, but there's some comparisons going on here that aren't quite as clear-cut as they seem at first glance.

    Figure 1 - Worldwide, 1999 Client and Server Operating Environment Revenues by Platform ($B)

    Well, DUH! Ain't much revenue for an "Operating Environment" that can be downloaded for free, so no wonder Linux lags behind 32-bit windows.

    Figure 2- Worldwide Client Operating Environment New License Shipment Shares 1999 and Shipment Growth 1999-2004

    Now that's better, as they are now comparing the number of Licenses instead of Dollars, but what do they define as a "client"? Does a TiVObox running Linux count? What about an IBM watch? Besides, I can just as well install the server version of Linux (or NT, for that matter) on my home PC.

    Figure 3 - Worldwide Server Operating Environment New License Shipment Shares 1999 and Shipment Growth 1999-2004

    Again, how do they define and differentiate between servers? I'd be willing to bet that a license for Solaris on a big Sun box is not really on the same par as a 486DX66 running a Linux server, but it seems that a license is a license is a license according to these stats.